DOCSLIB.ORG

Quantum Machine Learning for Classical Data

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Quantum Machine Learning for Classical Data Quantum Machine Learning For Classical Data Leonard P. Wossnig A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy of University College London. arXiv:2105.03684v2 [quant-ph] 12 May 2021 Department of Computer Science University College London May 13, 2021 2 I, Leonard P. Wossnig, confirm that the work presented in this thesis is my own. Where information has been derived from other sources, I confirm that this has been indicated in the work. Abstract In this dissertation, we study the intersection of quantum computing and supervised machine learning algorithms, which means that we investigate quantum algorithms for supervised machine learning that operate on classical data. This area of re- search falls under the umbrella of quantum machine learning, a research area of computer science which has recently received wide attention. In particular, we in- vestigate to what extent quantum computers can be used to accelerate supervised machine learning algorithms. The aim of this is to develop a clear understanding of the promises and limitations of the current state-of-the-art of quantum algorithms for supervised machine learning, but also to define directions for future research in this exciting field. We start by looking at supervised quantum machine learning (QML) algorithms through the lens of statistical learning theory. In this frame- work, we derive novel bounds on the computational complexities of a large set of supervised QML algorithms under the requirement of optimal learning rates. Next, we give a new bound for Hamiltonian simulation of dense Hamiltonians, a major subroutine of most known supervised QML algorithms, and then derive a classical algorithm with nearly the same complexity. We then draw the parallels to recent ‘quantum-inspired’ results, and will explain the implications of these results for quantum machine learning applications. Looking for areas which might bear larger advantages for QML algorithms, we finally propose a novel algorithm for Quantum Boltzmann machines, and argue that quantum algorithms for quantum data are one of the most promising applications for QML with potentially exponential advantage over classical approaches. Acknowledgements I want to thank foremost my supervisor and friend Simone Severini, who has always given me the freedom to pursue any direction I found interesting and promising, and has served me as a guide through most of my PhD. Next, I want to thank Aram Harrow, my secondary advisor, who has always been readily available to answer my questions and discuss a variety of research topics with me. I also want to thank Carlo Ciliberto, Nathan Wiebe, and Patrick Rebentrost, who have worked closely with me and have also taught me most of the mathematical tricks and methods upon which my thesis is built. I furthermore want to thank all my collaborators throughout the years. These are in particular Chunhao Wang, Andrea Rocchetto, Marcello Benedetti, Alessan- dro Rudi, Raban Iten, Mark Herbster, Massimiliano Pontil, Maria Schuld, Zhikuan Zhao, Anupam Prakash, Shuxiang Cao, Hongxiang Chen, Shashanka Ubaru, Haim Avron, and Ivan Rungger. Almost in an equal contribution, I also want to thank Fernando Brandao,˜ Youssef Mroueh, Guang Hao Low, Robin Kothari, Yuan Su, Tongyang Li, Ewin Tang, Kanav Setia, Matthias Troyer, and Damian Steiger for many helpful discussions, feedback, and enlightening explanations. I am particularly grateful to Edward Grant, Miriam Cha, and Ian Horobin, who made it possible for me to write this thesis. I want to acknowledge UCL for giving me the opportunity to pursue this PhD thesis, and acknowledge the kind support of align Royal Society Research grant and the Google PhD Fellowship, which gave me the freedom to work on these interesting topics. Acknowledgements 5 Portions of the work that are included in this thesis were completed while I was visiting the Institut Henri Poincare´ of the Sorbonne University in Paris. I particularly want to thank Riam Kim-McLeod for the support and help with the editing of the thesis. I finally want to thank my family for the continued love and support. Impact Statement Quantum machine learning bears promises for many areas, ranging from the health- care to the financial industry. In today’s world, where data is available in abundance, only novel algorithms and approaches are enabling us to make reliable predictions that can enhance our life, productivity, or wealth. While Moore’s law is coming to an end, novel computational paradigms are sought after to enable a further growth of processing power. Quantum computing has become one of the prominent can- didates, and is maturing rapidly. The here presented PhD thesis develops and stud- ies this novel computational paradigm in light of existing classical solutions and thereby develops a path towards quantum algorithms that can outperform classical approaches. Contents 1 Introduction and Overview 12 1.1 Synopsis of the thesis . 15 1.2 Summary of our contributions . 16 1.3 Statement of authorship . 20 2 Notation And Mathematical Preliminaries 23 2.1 Notation . 23 2.2 Matrix functional analysis . 24 3 Statistical Learning Theory 29 3.1 Review of key results in Learning Theory . 31 3.1.1 Supervised Learning . 32 3.1.2 Empirical risk minimization and learning rates . 33 3.1.3 Regularisation and modern approaches . 37 3.2 Review of supervised quantum machine learning algorithms . 39 3.2.1 Recap: Quantum Linear Regression and Least Squares . 40 3.2.2 Recap: Quantum Support Vector Machine . 43 3.3 Analysis of quantum machine learning algorithms . 44 3.3.1 Bound on the optimisation error . 46 3.3.2 Bounds on the sampling error . 49 3.3.3 Bounds on the condition number . 50 3.4 Analysis of supervised QML algorithms . 57 3.5 Conclusion . 58 Contents 8 4 randomised Numerical Linear Algebra 60 4.1 Introduction . 61 4.2 Memory models and memory access . 62 4.2.1 The pass efficient model . 62 4.2.2 Quantum random access memory . 64 4.2.3 Quantum inspired memory structures . 67 4.3 Basic matrix multiplication . 69 4.4 Hamiltonian Simulation . 74 4.4.1 Introduction . 75 4.4.2 Related work . 80 4.4.3 Applications . 85 4.4.4 Hamiltonian Simulation for dense matrices . 87 4.4.5 Hamiltonian Simulation with the Nystrom¨ method . 102 4.4.6 Beyond the Nystrom¨ method . 126 4.4.7 Conclusion . 126 5 Promising avenues for QML 130 5.1 Generative quantum machine learning . 131 5.1.1 Related work . 132 5.2 Boltzmann and quantum Boltzmann machines . 133 5.3 Training quantum Boltzmann machines . 135 5.3.1 Variational training for restricted Hamiltonians . 137 5.3.2 Gradient based training for general Hamiltonians . 143 5.4 Conclusion . 147 6 Conclusions 150 Appendices 156 A Appendix 1: Quantum Subroutines 156 A.1 Amplitude estimation . 156 A.2 The Hadamard test . 157 Contents 9 B Appendix 2: Deferred proofs 159 B.1 Derivation of the variational bound . 159 B.2 Gradient estimation . 161 B.2.1 Operationalizing the gradient based training . 163 B.3 Approach 2: Divided Differences . 170 B.3.1 Operationalising . 187 Bibliography 192 List of Figures 1.1 Different fields of study in Quantum Machine Learning. The dif- ferent areas are related to the choice of algorithm, i.e., whether it is executed on a quantum or classical computer, and the choice of the target problem, i.e., whether it operates on quantum or classical data. 13 3.1 Summary of time complexities for training and testing of differ- ent classical and quantum algorithms when statistical guarantees are taken into account. We omit polylog(n;d) dependencies for the p quantum algorithms. We assume e = Q(1= n) and count the ef- fects of measurement errors. The acronyms in the table refer to: least square support vector machines (LS-SVM), kernel ridge re- gression (KRR), quantum kernel least squares (QKLS), quantum kernel linear regression (QKLR), and quantum support vector ma- chines (QSVM). Note that for quantum algorithms the state ob- tained after training cannot be maintained or copied and the algo- rithm must be retrained after each test round. This brings a factor proportional to the train time in the test time of quantum algorithms. Because the condition number may also depend on n and for quan- tum algorithms this dependency may be worse, the overall scaling of the quantum algorithms may be slower than the classical. 59 4.1 An example of the data structure that allows for efficient state prepa- ration using a logarithmic number of conditional rotations. 66 List of Figures 11 4.2 An example of the classical (dynamic) data structure that enables × efficient sample and query access for the example of H 2 C2 4... 69 p 4.3 Comparing our result O(t d kHk polylog(t;d;kHk;1=e)) with other quantum and classical algorithms for different models. Since the qRAM model is stronger than the sparse-access model and the classical sampling and query access model, we consider the advan- tage of our algorithm against others when they are directly applied to the qRAM model. 86 5.1 Comparison of previous training algorithms for quantum Boltz- mann machines. The models have a varying cost function (objec- tive), contain (are able to be trained with) hidden units, and have different input data (classical or quantum). 132 Chapter 1 Introduction and Overview In the last twenty years, due to increased computational power and the availability of vast amounts of data, machine learning (ML) has seen an immense success, with applications ranging from computer vision [1] to playing complex games such as the Atari series [2] or the traditional game of Go [3]. However, over the past few years, challenges have surfaced that threaten the end of this revolution.
Load more

Recommended publications
  • Simulating Quantum Field Theory with a Quantum Computer
    Simulating quantum field theory with a quantum computer John Preskill Lattice 2018 28 July 2018 This talk has two parts (1) Near-term prospects for quantum computing. (2) Opportunities in quantum simulation of quantum field theory. Exascale digital computers will advance our knowledge of QCD, but some challenges will remain, especially concerning real-time evolution and properties of nuclear matter and quark-gluon plasma at nonzero temperature and chemical potential. Digital computers may never be able to address these (and other) problems; quantum computers will solve them eventually, though I’m not sure when. The physics payoff may still be far away, but today’s research can hasten the arrival of a new era in which quantum simulation fuels progress in fundamental physics. Frontiers of Physics short distance long distance complexity Higgs boson Large scale structure “More is different” Neutrino masses Cosmic microwave Many-body entanglement background Supersymmetry Phases of quantum Dark matter matter Quantum gravity Dark energy Quantum computing String theory Gravitational waves Quantum spacetime particle collision molecular chemistry entangled electrons A quantum computer can simulate efficiently any physical process that occurs in Nature. (Maybe. We don’t actually know for sure.) superconductor black hole early universe Two fundamental ideas (1) Quantum complexity Why we think quantum computing is powerful. (2) Quantum error correction Why we think quantum computing is scalable. A complete description of a typical quantum state of just 300 qubits requires more bits than the number of atoms in the visible universe. Why we think quantum computing is powerful We know examples of problems that can be solved efficiently by a quantum computer, where we believe the problems are hard for classical computers.
    [Show full text]
  • Quantum Machine Learning: Benefits and Practical Examples
    Quantum Machine Learning: Benefits and Practical Examples Frank Phillipson1[0000-0003-4580-7521] 1 TNO, Anna van Buerenplein 1, 2595 DA Den Haag, The Netherlands [email protected] Abstract. A quantum computer that is useful in practice, is expected to be devel- oped in the next few years. An important application is expected to be machine learning, where benefits are expected on run time, capacity and learning effi- ciency. In this paper, these benefits are presented and for each benefit an example application is presented. A quantum hybrid Helmholtz machine use quantum sampling to improve run time, a quantum Hopfield neural network shows an im- proved capacity and a variational quantum circuit based neural network is ex- pected to deliver a higher learning efficiency. Keywords: Quantum Machine Learning, Quantum Computing, Near Future Quantum Applications. 1 Introduction Quantum computers make use of quantum-mechanical phenomena, such as superposi- tion and entanglement, to perform operations on data [1]. Where classical computers require the data to be encoded into binary digits (bits), each of which is always in one of two definite states (0 or 1), quantum computation uses quantum bits, which can be in superpositions of states. These computers would theoretically be able to solve certain problems much more quickly than any classical computer that use even the best cur- rently known algorithms. Examples are integer factorization using Shor's algorithm or the simulation of quantum many-body systems. This benefit is also called ‘quantum supremacy’ [2], which only recently has been claimed for the first time [3]. There are two different quantum computing paradigms.
    [Show full text]
  • COVID-19 Detection on IBM Quantum Computer with Classical-Quantum Transfer Learning
    medRxiv preprint doi: https://doi.org/10.1101/2020.11.07.20227306; this version posted November 10, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license . Turk J Elec Eng & Comp Sci () : { © TUB¨ ITAK_ doi:10.3906/elk- COVID-19 detection on IBM quantum computer with classical-quantum transfer learning Erdi ACAR1*, Ihsan_ YILMAZ2 1Department of Computer Engineering, Institute of Science, C¸anakkale Onsekiz Mart University, C¸anakkale, Turkey 2Department of Computer Engineering, Faculty of Engineering, C¸anakkale Onsekiz Mart University, C¸anakkale, Turkey Received: .201 Accepted/Published Online: .201 Final Version: ..201 Abstract: Diagnose the infected patient as soon as possible in the coronavirus 2019 (COVID-19) outbreak which is declared as a pandemic by the world health organization (WHO) is extremely important. Experts recommend CT imaging as a diagnostic tool because of the weak points of the nucleic acid amplification test (NAAT). In this study, the detection of COVID-19 from CT images, which give the most accurate response in a short time, was investigated in the classical computer and firstly in quantum computers. Using the quantum transfer learning method, we experimentally perform COVID-19 detection in different quantum real processors (IBMQx2, IBMQ-London and IBMQ-Rome) of IBM, as well as in different simulators (Pennylane, Qiskit-Aer and Cirq). By using a small number of data sets such as 126 COVID-19 and 100 Normal CT images, we obtained a positive or negative classification of COVID-19 with 90% success in classical computers, while we achieved a high success rate of 94-100% in quantum computers.
    [Show full text]
  • Quantum Inductive Learning and Quantum Logic Synthesis
    Portland State University PDXScholar Dissertations and Theses Dissertations and Theses 2009 Quantum Inductive Learning and Quantum Logic Synthesis Martin Lukac Portland State University Follow this and additional works at: https://pdxscholar.library.pdx.edu/open_access_etds Part of the Electrical and Computer Engineering Commons Let us know how access to this document benefits ou.y Recommended Citation Lukac, Martin, "Quantum Inductive Learning and Quantum Logic Synthesis" (2009). Dissertations and Theses. Paper 2319. https://doi.org/10.15760/etd.2316 This Dissertation is brought to you for free and open access. It has been accepted for inclusion in Dissertations and Theses by an authorized administrator of PDXScholar. For more information, please contact [email protected]. QUANTUM INDUCTIVE LEARNING AND QUANTUM LOGIC SYNTHESIS by MARTIN LUKAC A dissertation submitted in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY in ELECTRICAL AND COMPUTER ENGINEERING. Portland State University 2009 DISSERTATION APPROVAL The abstract and dissertation of Martin Lukac for the Doctor of Philosophy in Electrical and Computer Engineering were presented January 9, 2009, and accepted by the dissertation committee and the doctoral program. COMMITTEE APPROVALS: Irek Perkowski, Chair GarrisoH-Xireenwood -George ^Lendaris 5artM ?teven Bleiler Representative of the Office of Graduate Studies DOCTORAL PROGRAM APPROVAL: Malgorza /ska-Jeske7~Director Electrical Computer Engineering Ph.D. Program ABSTRACT An abstract of the dissertation of Martin Lukac for the Doctor of Philosophy in Electrical and Computer Engineering presented January 9, 2009. Title: Quantum Inductive Learning and Quantum Logic Synhesis Since Quantum Computer is almost realizable on large scale and Quantum Technology is one of the main solutions to the Moore Limit, Quantum Logic Synthesis (QLS) has become a required theory and tool for designing Quantum Logic Circuits.
    [Show full text]
  • Nearest Centroid Classification on a Trapped Ion Quantum Computer
    www.nature.com/npjqi ARTICLE OPEN Nearest centroid classification on a trapped ion quantum computer ✉ Sonika Johri1 , Shantanu Debnath1, Avinash Mocherla2,3,4, Alexandros SINGK2,3,5, Anupam Prakash2,3, Jungsang Kim1 and Iordanis Kerenidis2,3,6 Quantum machine learning has seen considerable theoretical and practical developments in recent years and has become a promising area for finding real world applications of quantum computers. In pursuit of this goal, here we combine state-of-the-art algorithms and quantum hardware to provide an experimental demonstration of a quantum machine learning application with provable guarantees for its performance and efficiency. In particular, we design a quantum Nearest Centroid classifier, using techniques for efficiently loading classical data into quantum states and performing distance estimations, and experimentally demonstrate it on a 11-qubit trapped-ion quantum machine, matching the accuracy of classical nearest centroid classifiers for the MNIST handwritten digits dataset and achieving up to 100% accuracy for 8-dimensional synthetic data. npj Quantum Information (2021) 7:122 ; https://doi.org/10.1038/s41534-021-00456-5 INTRODUCTION Thus, one might hope that noisy quantum computers are 1234567890():,; Quantum technologies promise to revolutionize the future of inherently better suited for machine learning computations than information and communication, in the form of quantum for other types of problems that need precise computations like computing devices able to communicate and process massive factoring or search problems. amounts of data both efficiently and securely using quantum However, there are significant challenges to be overcome to resources. Tremendous progress is continuously being made both make QML practical.
    [Show full text]
  • High Energy Physics Quantum Computing
    High Energy Physics Quantum Computing Quantum Information Science in High Energy Physics at the Large Hadron Collider PI: O.K. Baker, Yale University Unraveling the quantum structure of QCD in parton shower Monte Carlo generators PI: Christian Bauer, Lawrence Berkeley National Laboratory Co-PIs: Wibe de Jong and Ben Nachman (LBNL) The HEP.QPR Project: Quantum Pattern Recognition for Charged Particle Tracking PI: Heather Gray, Lawrence Berkeley National Laboratory Co-PIs: Wahid Bhimji, Paolo Calafiura, Steve Farrell, Wim Lavrijsen, Lucy Linder, Illya Shapoval (LBNL) Neutrino-Nucleus Scattering on a Quantum Computer PI: Rajan Gupta, Los Alamos National Laboratory Co-PIs: Joseph Carlson (LANL); Alessandro Roggero (UW), Gabriel Purdue (FNAL) Particle Track Pattern Recognition via Content-Addressable Memory and Adiabatic Quantum Optimization PI: Lauren Ice, Johns Hopkins University Co-PIs: Gregory Quiroz (Johns Hopkins); Travis Humble (Oak Ridge National Laboratory) Towards practical quantum simulation for High Energy Physics PI: Peter Love, Tufts University Co-PIs: Gary Goldstein, Hugo Beauchemin (Tufts) High Energy Physics (HEP) ML and Optimization Go Quantum PI: Gabriel Perdue, Fermilab Co-PIs: Jim Kowalkowski, Stephen Mrenna, Brian Nord, Aris Tsaris (Fermilab); Travis Humble, Alex McCaskey (Oak Ridge National Lab) Quantum Machine Learning and Quantum Computation Frameworks for HEP (QMLQCF) PI: M. Spiropulu, California Institute of Technology Co-PIs: Panagiotis Spentzouris (Fermilab), Daniel Lidar (USC), Seth Lloyd (MIT) Quantum Algorithms for Collider Physics PI: Jesse Thaler, Massachusetts Institute of Technology Co-PI: Aram Harrow, Massachusetts Institute of Technology Quantum Machine Learning for Lattice QCD PI: Boram Yoon, Los Alamos National Laboratory Co-PIs: Nga T. T. Nguyen, Garrett Kenyon, Tanmoy Bhattacharya and Rajan Gupta (LANL) Quantum Information Science in High Energy Physics at the Large Hadron Collider O.K.
    [Show full text]
  • Modeling Observers As Physical Systems Representing the World from Within: Quantum Theory As a Physical and Self-Referential Theory of Inference
    Modeling observers as physical systems representing the world from within: Quantum theory as a physical and self-referential theory of inference John Realpe-G´omez1∗ Theoretical Physics Group, School of Physics and Astronomy, The University of Manchestery, Manchester M13 9PL, United Kingdom and Instituto de Matem´aticas Aplicadas, Universidad de Cartagena, Bol´ıvar130001, Colombia (Dated: June 13, 2019) In 1929 Szilard pointed out that the physics of the observer may play a role in the analysis of experiments. The same year, Bohr pointed out that complementarity appears to arise naturally in psychology where both the objects of perception and the perceiving subject belong to `our mental content'. Here we argue that the formalism of quantum theory can be derived from two related intu- itive principles: (i) inference is a classical physical process performed by classical physical systems, observers, which are part of the experimental setup|this implies non-commutativity and imaginary- time quantum mechanics; (ii) experiments must be described from a first-person perspective|this leads to self-reference, complementarity, and a quantum dynamics that is the iterative construction of the observer's subjective state. This approach suggests a natural explanation for the origin of Planck's constant as due to the physical interactions supporting the observer's information process- ing, and sheds new light on some conceptual issues associated to the foundations of quantum theory. It also suggests that fundamental equations in physics are typically
    [Show full text]
  • Quantum Computing Methods for Supervised Learning Arxiv
    Quantum Computing Methods for Supervised Learning Viraj Kulkarni1, Milind Kulkarni1, Aniruddha Pant2 1 Vishwakarma University 2 DeepTek Inc June 23, 2020 Abstract The last two decades have seen an explosive growth in the theory and practice of both quantum computing and machine learning. Modern machine learning systems process huge volumes of data and demand massive computational power. As silicon semiconductor miniaturization approaches its physics limits, quantum computing is increasingly being considered to cater to these computational needs in the future. Small-scale quantum computers and quantum annealers have been built and are already being sold commercially. Quantum computers can benefit machine learning research and application across all science and engineering domains. However, owing to its roots in quantum mechanics, research in this field has so far been confined within the purview of the physics community, and most work is not easily accessible to researchers from other disciplines. In this paper, we provide a background and summarize key results of quantum computing before exploring its application to supervised machine learning problems. By eschewing results from physics that have little bearing on quantum computation, we hope to make this introduction accessible to data scientists, machine learning practitioners, and researchers from across disciplines. 1 Introduction Supervised learning is the most commonly applied form of machine learning. It works in two arXiv:2006.12025v1 [quant-ph] 22 Jun 2020 stages. During the training stage, the algorithm extracts patterns from the training dataset that contains pairs of samples and labels and converts these patterns into a mathematical representation called a model. During the inference stage, this model is used to make predictions about unseen samples.
    [Show full text]
  • High Energy Physics Quantum Information Science Awards Abstracts
    High Energy Physics Quantum Information Science Awards Abstracts Towards Directional Detection of WIMP Dark Matter using Spectroscopy of Quantum Defects in Diamond Ronald Walsworth, David Phillips, and Alexander Sushkov Challenges and Opportunities in Noise‐Aware Implementations of Quantum Field Theories on Near‐Term Quantum Computing Hardware Raphael Pooser, Patrick Dreher, and Lex Kemper Quantum Sensors for Wide Band Axion Dark Matter Detection Peter S Barry, Andrew Sonnenschein, Clarence Chang, Jiansong Gao, Steve Kuhlmann, Noah Kurinsky, and Joel Ullom The Dark Matter Radio‐: A Quantum‐Enhanced Dark Matter Search Kent Irwin and Peter Graham Quantum Sensors for Light-field Dark Matter Searches Kent Irwin, Peter Graham, Alexander Sushkov, Dmitry Budke, and Derek Kimball The Geometry and Flow of Quantum Information: From Quantum Gravity to Quantum Technology Raphael Bousso1, Ehud Altman1, Ning Bao1, Patrick Hayden, Christopher Monroe, Yasunori Nomura1, Xiao‐Liang Qi, Monika Schleier‐Smith, Brian Swingle3, Norman Yao1, and Michael Zaletel Algebraic Approach Towards Quantum Information in Quantum Field Theory and Holography Daniel Harlow, Aram Harrow and Hong Liu Interplay of Quantum Information, Thermodynamics, and Gravity in the Early Universe Nishant Agarwal, Adolfo del Campo, Archana Kamal, and Sarah Shandera Quantum Computing for Neutrino‐nucleus Dynamics Joseph Carlson, Rajan Gupta, Andy C.N. Li, Gabriel Perdue, and Alessandro Roggero Quantum‐Enhanced Metrology with Trapped Ions for Fundamental Physics Salman Habib, Kaifeng Cui1,
    [Show full text]
  • Michael Bremner T + 44 1173315236 B [email protected] Skype: Mickbremner
    Department of Computer Science University of Bristol Woodland Road BS8 1UB Bristol United Kingdom H +44 7887905572 Michael Bremner T + 44 1173315236 B [email protected] Skype: mickbremner Research Interests: Quantum simulation, computational complexity, quantum computing architectures, fault tolerance in quantum computing, and quantum control theory. Personal details Date of birth: 5th August 1978 Languages: English Nationality: Australian Marital status: Unmarried Education 2001–2005 PhD, Department of Physics, University of Queensland. Project: Characterizing entangling quantum dynamics. Supervised by Prof. Michael Nielsen and Prof. Gerard Milburn. 2000 BSc with Honours Class I, Department of Physics, University of Queensland. Project: Entanglement generation and tests of local realism in quantum optics. Supervised by Prof. Tim Ralph and Dr Bill Munro. 1997–1999 BSc (Physics), University of Queensland. 1996 Senior certificate, Villanova College, Brisbane. Completed high school education. Professional experience 2007–Present Postdoctoral researcher, Department of Computer Science, University of Bristol. Supervised by Prof. Richard Jozsa. 2005–2007 Postdoctoral researcher, Institute for Theoretical Physics and Institute for Quantum Optics and Quantum Information, University of Innsbruck. Supervised by Prof. Hans Briegel Scholarships and awards 2005 Dean’s commendation for excellence in a PhD thesis, University of Queensland. 2001–2005 Australian Postgraduate Award, Australian Research Council. 1998–1999 Scholarships for summer vacation research at the University of Queensland 1997–1999 Four Dean’s commendations for high achievement, University of Queensland. 1996 Australian Students Prize – Awarded to the top 500 students completing high school studies. 1/5 Recent conference presentations 2009 D. Shepherd (speaker) and M. J. Bremner Instantaneous Quantum Computation con- tributed talk at Quantum Information Processing 2009, Santa Fe.
    [Show full text]
  • Federated Quantum Machine Learning
    entropy Article Federated Quantum Machine Learning Samuel Yen-Chi Chen * and Shinjae Yoo Computational Science Initiative, Brookhaven National Laboratory, Upton, NY 11973, USA; [email protected] * Correspondence: [email protected] Abstract: Distributed training across several quantum computers could significantly improve the training time and if we could share the learned model, not the data, it could potentially improve the data privacy as the training would happen where the data is located. One of the potential schemes to achieve this property is the federated learning (FL), which consists of several clients or local nodes learning on their own data and a central node to aggregate the models collected from those local nodes. However, to the best of our knowledge, no work has been done in quantum machine learning (QML) in federation setting yet. In this work, we present the federated training on hybrid quantum-classical machine learning models although our framework could be generalized to pure quantum machine learning model. Specifically, we consider the quantum neural network (QNN) coupled with classical pre-trained convolutional model. Our distributed federated learning scheme demonstrated almost the same level of trained model accuracies and yet significantly faster distributed training. It demonstrates a promising future research direction for scaling and privacy aspects. Keywords: quantum machine learning; federated learning; quantum neural networks; variational quantum circuits; privacy-preserving AI Citation: Chen, S.Y.-C.; Yoo, S. 1. Introduction Federated Quantum Machine Recently, advances in machine learning (ML), in particular deep learning (DL), have Learning. Entropy 2021, 23, 460. found significant success in a wide variety of challenging tasks such as computer vi- https://doi.org/10.3390/e23040460 sion [1–3], natural language processing [4], and even playing the game of Go with a superhuman performance [5].
    [Show full text]
  • Future Directions of Quantum Information Processing a Workshop on the Emerging Science and Technology of Quantum Computation, Communication, and Measurement
    Future Directions of Quantum Information Processing A Workshop on the Emerging Science and Technology of Quantum Computation, Communication, and Measurement Seth Lloyd, Massachusetts Institute of Technology Dirk Englund, Massachusetts Institute of Technology Workshop funded by the Basic Research Office, Office of the Assistant Prepared by Kate Klemic Ph.D. and Jeremy Zeigler Secretary of Defense for Research & Engineering. This report does not Virginia Tech Applied Research Corporation necessarily reflect the policies or positions of the US Department of Defense Preface Over the past century, science and technology have brought remarkable new capabilities to all sectors of the economy; from telecommunications, energy, and electronics to medicine, transportation and defense. Technologies that were fantasy decades ago, such as the internet and mobile devices, now inform the way we live, work, and interact with our environment. Key to this technological progress is the capacity of the global basic research community to create new knowledge and to develop new insights in science, technology, and engineering. Understanding the trajectories of this fundamental research, within the context of global challenges, empowers stakeholders to identify and seize potential opportunities. The Future Directions Workshop series, sponsored by the Basic Research Office of the Office of the Assistant Secretary of Defense for Research and Engineering, seeks to examine emerging research and engineering areas that are most likely to transform future technology capabilities. These workshops gather distinguished academic and industry researchers from the world’s top research institutions to engage in an interactive dialogue about the promises and challenges of these emerging basic research areas and how they could impact future capabilities.
    [Show full text]